Regionality Revisited: An Examination of the Direction of Spread of Currency Crises∗

Currency crises appear to spread regionally. In this paper we use Bayesian methodology on data from four waves of currency crises in the late twentieth century to examine the mechanism by which currency crises spread from one country to another. In particular, we examine whether trade linkages, “other” unquantifiable linkages proxied for by relative geographical distance, or macroeconomic and financial similarity are important (separately and together) in determining the direction of spread of currency crises. Our results on trade links represent a Bayesian reworking of the results presented in Glick and Rose (1998). Our examination of other forms of linkages extends and sharpens their analysis. Our results suggest that trade linkages are consistently important in determining the spread of currency crises, as are, to a lesser extent, macroeconomic and financial similarity. However, we find that in comparison geographical distance is not a good indicator for the spread of crises. In this paper we apply some relatively recent techniques in the Bayesian analysis of binary data that allow us to overcome small sample biases and meaningfully collate lessons across waves of currency crises. The econometric methodology highlighted in this paper is hitherto unused in the empirical literature on currency crises and bank runs and is widely applicable to the same. ∗I thank John Rust and Giancarlo Corsetti for their advice. I am grateful to Hugo Beńıtez-Silva for his help with Gauss and for general encouragement and support. †Department of Economics, Yale University, P.O.Box 208268, New Haven, CT 06520-8268, USA. E-mail: [email protected], http://www.econ.yale.edu/~amil